Using Semi-Structured Interviews to Better Understand Nonresponse to a Government-Sponsored, Mail-Based Survey of Households
Using Semi-Structured Interviews to Better Understand Nonresponse to a Government-Sponsored, Mail-Based Survey of Households (DOI: https://doi.org/10.1093/poq/nfaf026) was written by Rebecca L Medway, Melissa Scardaville, Paula Dias, and Christopher Paek in 2025. It was published in Public Opinion Quarterly (vol. 89).
Between 2012 and 2023, response rate (AAPOR RR3) for the NHES declined from 74% to 61%. To understand the causes of nonresponse, AIR conducted in-person interviews with screener non-respondents in the 2019 administration.
Initial sampling frame is the final reminder mail file (i.e., with respondents cut). Further removed hard refusals, addresses that returned undeliverable nixie codes, P.O. boxes, and addresses on rural mail routes. Finally, also restricted to addresses within 30 miles of 4 specific sites; target sample is 100 households from each site for a total of 400.
- oversample addresses that were believed to include children, Hispanic household members, or a head of household with lower educational attainment
- no within-household sampling; any eligible household member is allowed to participate
- eligible meaning 18+ and resident from January to April 2019
Details of administration:
- fielded April to June 2019
- contact by mail, phone, and in-person visits
- bilingual recruitment (English and Spanish)
- invite included a $5 incentive
- $120 incentive for completing an interview
AIR completed 85 interviews; response rate (AAPOR RR1) was 22%. 15 completed the screener after being sampled for an interview, so are excluded from analysis.
Analysis is qualitative using modified grounded theory approach. Key findings:
- objective and subjective time use considerations are related to nonresponse
government label carries trust but prospective respondents do perceive a cost in terms of privacy from specific requests, and do believe that their participation does not matter (i.e., assume the government already knows)
- privacy concerns driven by history of data breaches/leaks; do not trust researcher assurances
- "Emphasizing data security for those concerned about breaches may be interpreted as disingenuous by those who hold the incorrect belief that there are centralized data repositories that all government agencies can access."
Changes made to NCES as a result include:
- emphasize deadline to respond
- refine FAQ to better address concerns
- highlight data security policies of sponsor on the study website